U.S. patent number 6,930,637 [Application Number 10/293,834] was granted by the patent office on 2005-08-16 for method and apparatus for high resolution tracking via mono-pulse beam-forming in a communication system.
This patent grant is currently assigned to Texas Instruments Incorporated. Invention is credited to Louis R. Brothers, Jr., John Cangeme, Alexander Flaig, Samuel J. MacMullen, H. Vincent Poor, Tandhoni S. Rao, Stuart C. Schwartz, Triveni N. Upadhyay.
United States Patent |
6,930,637 |
Brothers, Jr. , et
al. |
August 16, 2005 |
Method and apparatus for high resolution tracking via mono-pulse
beam-forming in a communication system
Abstract
Method and apparatus for high resolution tracking via mono-pulse
beam-forming in a communication system in which the capacity and
range of mobile or fixed wireless communication base stations are
improved by implementing a single or multiple antenna beam per
signal path. Adaptive beam-forming based on up-link direction-of
arrival estimation can be performed without using the
above-mentioned computationally intensive techniques.
Inventors: |
Brothers, Jr.; Louis R.
(Dorchester, MA), Cangeme; John (Billerica, MA), Flaig;
Alexander (Concord, MA), MacMullen; Samuel J. (Carlisle,
MA), Poor; H. Vincent (Princeton, NJ), Rao; Tandhoni
S. (Ashland, MA), Schwartz; Stuart C. (Princeton,
NJ), Upadhyay; Triveni N. (Concord, MA) |
Assignee: |
Texas Instruments Incorporated
(Dallas, TX)
|
Family
ID: |
31996872 |
Appl.
No.: |
10/293,834 |
Filed: |
November 13, 2002 |
Current U.S.
Class: |
342/427; 342/377;
455/561 |
Current CPC
Class: |
G01S
3/32 (20130101); H01Q 25/02 (20130101); H04B
7/0617 (20130101); H04B 7/086 (20130101) |
Current International
Class: |
G01S
3/32 (20060101); G01S 3/14 (20060101); H01Q
25/02 (20060101); H01Q 25/00 (20060101); H04B
7/04 (20060101); H04B 7/08 (20060101); H04B
7/06 (20060101); G01S 005/02 (); H01Q 003/16 () |
Field of
Search: |
;342/377,383,427
;455/561,562.1 |
References Cited
[Referenced By]
U.S. Patent Documents
Foreign Patent Documents
Other References
Lo, K. W. "Adaptivity of a Real Symmetric Array by DOA Estimation
and Null Steering," IEE Proc-Padar, Sonar, Navigation, vol. 144 No.
5, Oct. 1997, pp. 245-251. .
Morrixon, Andrew et al, "A Space-Time Beamforming RAKE Receiver for
Third Generation Wideband CDMA Base Stations," ICCE International
Conference on Consumer Electronics, Jun. 1999, pp. 312-313. .
Winters, Jack H., "Smart Antennas for Wireless Systems," IEEE
Personal Communications, Feb. 1998, pp. 23-27. .
Kuchar, Alexander, et al, "Real-Time Smart Antenna Processing for
GSM1800 Base Station," IEEE 49th Vehicular Technology Conference,
May 1999, pp. 664-669 vol. 1. .
Kederer, Werner et al, "Direction of Arrival (DOA) Determination
Based on Monopulse Concepts," 2000 Asia Pacific Microwave
Conference, Dec. 2000, pp. 120-123..quadrature..quadrature.. .
Sheng, W.X. et al, "Super-Resolution DOA Estimation in Switch Beam
Smart Antenna," ISAPE 2000, 5th International Symposium on
Antennas, Propagation and EM Theory, Aug. 2000, pp. 603-606. .
Kederer, Werner et al, "Direction of Arrival (DOA) Determination
Based on Monopulse COcepts," 2000 Asia-Pacific Microwav Conference,
Dec. 2000, pp. 120-123. .
Lo, K. W., "Adaptivity of a Real-Symmetric Array by DOA Estimation
and NullSteering," IEE Proceedings Radar, Sonar, and Navigation,
Oct. 1997, pp. 245-251. .
"Smart Antennas For Mobile Communication Systems: Benefits and
Challenges", G.V. Tsoulos, Electronics & Communication
Engineering Journal, Apr. 1999, pp. 84-94. .
"Smart Antennas", Michael Chryssomailis, IEEE Antennas and
Propagation Magazine, vol. 42, No. 3, Jun. 2000, pp. 129-136. .
"A Comparison of Tracking-Beam Arrays and Switching-Beam Arrays
Operating in a CDMA Mobile Communication Channel", IEEE Antennas
and Propagation Magazine, vol. 41, No. 6, Dec. 1999, pp.
10-22..
|
Primary Examiner: Issing; Gregory C.
Attorney, Agent or Firm: Neerings; Ronald O. Brady, III;
Wade James Telecky, Jr.; Frederick J.
Parent Case Text
This application claims priority under 35 USC .sctn.119(e)(1) of
provisional application number 60/331,423, filed Nov. 15, 2001.
Claims
What is claimed:
1. A method of tracking a selected radio-frequency signal in a
multiple access communication system, comprising the steps of:
determining an estimated direction of arrival of a signal received
at an antenna location; demodulating a received signal, from the
estimated direction of arrival, into a digital base-band signal;
descrambling the digital base-band signal; separating a selected
signal from co-channel interference in the descrambled signal;
tracking the direction of arrival of the selected signal by
performing a sequence of steps comprising: measuring a summation
beam at the estimated direction of arrival; measuring a difference
beam at the estimated direction of arrival; comparing the summation
and difference beams to derive an angular offset; updating the
direction of arrival by applying the angular offset to the
estimated direction of arrival; and repeating the measuring,
comparing, and updating steps.
2. The method of claim 1, wherein the step of determining an
estimated direction of arrival comprises determining the
orientation of a fixed location from which the signal is
transmitted relative to the antenna location.
3. The method of claim 1, wherein the step of determining an
estimated direction of arrival comprises: performing fixed-beam
searching over a sector extending from the antenna location.
4. The method of claim 3, wherein the step of performing fixed-beam
searching is performed substantially continuously as a background
process.
5. The method of claim 4, wherein the step of performing fixed-beam
searching comprises: receiving signals at a plurality of antennae
at the antenna location; performing a spatial transform of the
received signals at the plurality of antennae to derive a weight
vector.
6. The method of claim 5, wherein the step of measuring the
summation beam at the estimated direction of arrival comprises
summing signals from at least some of the plurality of antennae
using the weight vector; wherein the step of measuring the
difference beam at the estimated direction of arrival comprises
summing signals from at least some of the plurality of antennae in
a first half of an antenna array with opposite-phase signals from
at least some of the plurality of antennae in a second half of the
antenna array, using the weight vector.
7. The method of claim 6, wherein the first and second halves of
the antenna are first and second azimuth halves.
8. The method of claim 7, further comprising: summing signals from
at least some of the plurality of antennae in a top half of the
antennal array with opposite-phase signals from at least some of
the plurality of antennae in a bottom half of the antenna array,
using the weight vector.
9. The method of claim 1, wherein the signal comprises a
code-division-multiple-access (CDMA) encoded signal; and wherein
the separating step comprises despreading the CDMA encoded
signal.
10. The method of claim 9, further comprising: prior to the
tracking step, applying a match filter to the received signal.
11. The method of claim 1, wherein the signal comprises a
time-division-multiple-access (TDMA) encoded signal; and wherein
the separating step comprises deinterleaving the TDMA encoded
signal.
12. The method of claim 11, further comprising: prior to the
tracking step, applying a match filter to the received signal; and
also prior to the tracking step, applying a delay to the received
signal for synchronization.
13. A base station system for multiple access communication,
comprising: an antenna array; a plurality of radio frequency
demodulators for demodulating signals received at the antenna array
into spatial samples; a user processing element for separating
signals for a selected user from received signals from other users;
and a mono-pulse beam forming processor, comprising: a summing beam
former, for generating a summation signal from the plurality of
spatial samples for the selected user and a weighting vector; a
difference beam former, for generating a difference signal from the
plurality of spatial samples and a weighting vector; a channel
estimator, for generating an output signal corresponding to the
summation signal; and circuitry for generating an angle estimation
from the summation and difference signals and adjusting weighting
vectors responsive to the angle estimation.
14. The system of claim 13, further comprising: a spatial transform
processor, for transforming the plurality of spatial samples into a
plurality of beam signals for a plurality of users; a plurality of
user processing elements for determining a plurality of magnitudes
of spatial samples corresponding to each of the plurality of beam
signals; and a plurality of detectors for detecting the beam signal
having the highest magnitude, the weighting vectors corresponding
to ordered magnitudes of the beam signals.
15. The system of claim 13, wherein the antenna array is a linear
array of antennae.
16. The system of claim 13, wherein the antenna array is a
two-dimensional array of antennae.
17. The system of claim 13, wherein the plurality of radio
frequency demodulators demodulate code-division-multiple-access
(CDMA) signals received at the antenna array; and wherein the user
processing element separates signals by despreading the received
CDMA signals.
Description
BACKGROUND OF THE INVENTION
The invention relates to base station equipment for receiving and
transmitting one or more signals from one or more users, in which
the signals may arrive at the equipment along a multiplicity of
paths, from possibly different directions, and with possibly
different delays.
A major concern for providers of wireless communications services
is system coverage and capacity. Future systems promise data rates
and an aggregate capacity significantly higher than current
systems. However, with conventional base stations, the maximum link
closure range will be decreased substantially for users operating
at higher data rates. As a result, the promised data rates and
aggregate capacity can only be supported in a small region close to
the base station.
Smart antenna systems have been discussed in the literature as a
means of increasing capacity and coverage over and above that which
can be provided with simple omni-directional antennas. They achieve
this through spatio-temporal correlation of desired signals and
co-channel interference within a cell. Interference suppression is
implemented by forming narrow radiation patterns, forming radiation
nulls on significant interference points or a combination of the
two. Smart antennas are implemented in several forms;
switched-beam, Direction-of-Arrival (DOA) or Optimum Combining
derived adaptive-beams. Some systems are analog where the beam is
formed in an RF manifold such as a Butler matrix but the most
flexible are those that are digitally formed.
Switched beam systems such as the one described in U.S. Pat. No.
6,218,987 entitled "Radio Antenna System", form several fixed beams
in an RF Butler matrix with the ability to simultaneously broadcast
a common channel with a high gain wide beam. A similar fixed beam
system is described in U.S. Pat. No. 6,181,276 entitled "Sector
Shaping Transition System and Method" where a combination of a set
of fixed beams can be coherently combined, by analog means, to form
another beam that is better adapted to the area loading of the
cell. These switched-beam systems do not take advantage of the
maximum gain offered by the full aperture. As a mobile moves
through a cell it will suffer beam-width modulation as it travels
between the peaks of the several fixed beams. Furthermore,
switched-beam approaches simply further sub-divide a cell into
sub-sectors. Unfortunately, this method requires handoff between
the sub-sectors just as with a standard 3-sector system. These
handovers require valuable resources and ultimately reduce the
capacity of the network.
In the recent papers entitled "A comparison of Tracking-Beam arrays
and Switching-Beam Arrays Operating in a CDMA Mobile Communication
Channer", IEEE Antennas and Propagation Magazine, Dec. 1999, and
"Smart Antennas", IEEE Antennas and Propagation Magazine, Jun.
2000, "Smart Antennas for Mobile Communication systems: Benefits
and Challenges", Electronics and Communication Engineering Journal,
Apr. 1999, it has been shown that adaptive-beam systems perform
better than switched-beam systems especially in high interference
environments. They perform better partly because they take spatial
and temporal correlations of interfering signals into account and
eliminate the need for frequent handovers within a sector and also
tend to maintain maximum antenna gain in the desired direction
Adaptive-beam systems can be used to track individual mobile
terminals within a base station service area. Several different
methods are disclosed in the literature that all attempt to find an
array weighting vector that maximizes the SINR for a desired
signal. These methods vary in complexity. Estimation of signal
parameters by rotational invariance techniques (ESPRIT) is more
widely used than the other sub-space eigen-decomposition method
Multiple Signal Classification (MUSIC). Although MUSIC is
considered to achieve higher resolution it also requires more
computation in its searching algorithm than the closed form
solution provided by ESPRIT. These eigen-decomposition methods
require a good estimate of the array covariance matrix by averaging
over time, such that in the limit where the averaging time
approaches infinity the estimate becomes exact. The array
covariance matrix is found by averaging over several snapshots of
the array signal values. Once determined the matrix can be updated
after every sample. Several recent patent disclosures cite the use
of ESPRIT for adaptive beam-forming such as U.S. Pat. No. 6,008,759
entitled "Method of Determining the Direction of Arrival of a radio
Signal, as well as Radio Base Station and Radio Communications
Systems" and U.S. Pat. No. 5,892,700 "Method for the High
Resolution Evaluation of Signals for One or Two Dimensional
Directional or Frequency Estimation". In the preferred embodiment
of the former, a sub-optimal method is introduced that forms a beam
based on the steering vector determined from the strongest
eigen-value. This eliminates the need for full eigen-decomposition
and significantly reduces computation time allowing faster track
updates. It is considered sub-optimal because it does not attempt
to place nulls on significant interference and thus does not
maximizing SINR. However, it is suggested that interference may be
suppressed further by standard side-lobe control methods.
Furthermore, ubiquitous multi-path propagation with uncorrelated
fading would require at least two paths to be resolved requiring
means for several channels. In one embodiment of the latter, ESPRIT
is used to resolve several multi-path signals from a single desired
source simultaneously and therefore take advantage of maximal ratio
combining. Although this technique is considered optimum combining
because it maximizes SINR it has the disadvantage that its solution
does not necessarily place the peak of an antenna beam on the
desired signal path. The effect of this is the degradation in
sensitivity of the system to the thermal noise thus reducing the
range of the base station.
The methods described in the above-mentioned references require
computationally expensive eigen-decomposition of the estimated
array covariance matrix requiring at least an (M.times.M) matrix
inversion where M is the number of antenna elements. Accordingly as
the number of users, K, and the number of elements, M, grows the
matrix manipulation will become unwieldy and memory intensive.
Conversely, the method and apparatus of this invention replaces the
(M.times.M) matrix inversion to a single computation of a ratio.
Furthermore, the above-mentioned references describe methods that
incorporate switched and fixed beam solutions that carry the burden
of frequent handovers.
Other less computationally intensive methods for adaptive
beam-forming and direction finding do exist. For example, a simpler
means of DOA estimation disclosed in U.S. Pat. No. 6,212,406,
"Method for Providing Angular Diversity, and Base Station
Equipment", outlines a search and track by scan method, relying on
beam-width modulation, determines directions and delays of signals
by seeking the strongest power levels or largest SINR. In a
multi-path environment where the signal could jump discontinuously,
too much time could elapse before reacquiring the signal.
Furthermore, searching for a maximum signal to determine whether
maximum antenna gain has been achieved will prove difficult for
near-in high speed mobile units and signals experiencing fast
fading. Signal level measurement uncertainty could also be
construed as beam-width modulation further degrading accuracy.
Extensively used in many Radar and Sonar discriminators, another
successful technique utilized for DOA tracking is known as
mono-pulse beam forming and is described in "Introduction to Radar
Systems", M. I. Skolnik, 1980. Unlike ESPRIT and MUSIC techniques,
mono-pulse estimation of DOA requires only the determination of a
single ratio of two signals. The two signals are generated by
forming two different beams from a single antenna: 1) a summation
beam, 801, containing the signal information that is ultimately
carried through the rest of the network and 2) a difference beam,
802. Over angle space this ratio is a well behaved function, 901
from which an accurate estimate of DOA relative to the current beam
position may be determined. Its accuracy is an improvement over
beam peak finding because of the sharpness of the difference beam
null relative to the broad nature of the antenna beam. Its
performance does rely on the ability to find the zero of the
difference beam null and in a high interference or noisy
environment this null tends to fill increasing the uncertainty of
the angle offset estimate. Therefore, this technique requires a low
interference environment. However, mobile communication lends
itself to this technique due to the separation of radio links via
various multi-access schemes such as CDMA TDMA, and FDMA. Thus, low
interference is achieved through the orthogonality of the
co-channel users.
Historically, mono-pulse tracking, although simple to implement,
has not been utilized in multiple access communication systems.
Digital beam-forming has only recently started to make a presence
in practical systems due to the growth in processing speeds. Prior
to digital implementations beam-forming systems have typically been
realized in analog. To realize multiple beams in multiple access
systems would require separate analog channels in the antenna
beam-formers, including separate phase shifters and attenuators.
The number of phase shifters and attenuators could number in the
hundreds and even thousands per antenna, depending on the capacity
of the system and the number of antenna elements in the phased
array. This limits the number of simultaneous multiple beams to
tens-of-beams and not hundreds-of-beams required for multi-access
communication. Mono-pulse tracking has not been previously
implemented for this application because it implies the real-time
tracking of multiple simultaneous beams. Digital processors and
ASIC's have just recently surpassed the performance requirements to
achieve such a result. However, computational resources are still
and always will be considered premium. Thus, a need exists to
preserve as much of the computational resources as possible while
enabling a significant increase in the capacity and range of
communication systems.
SUMMARY OF THE INVENTION
In one embodiment of the invention the capacity and range of mobile
or fixed wireless communication base stations are improved by
implementing a single or multiple antenna beam per signal path.
Adaptive beam-forming based on up-link direction-of-arrival
estimation can be performed without using the above-mentioned
computationally intensive techniques.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 illustrates the low resolution angle estimation with a set
multiple fixed beams in a multi-user multi-path environment.
FIG. 2 depicts a snapshot of a set of agile beams prior to
track.
FIG. 3 depicts a snapshot of a set of agile beams positioned at
locations determined by the fixed beam search processor.
FIG. 4 depicts a snapshot of a set of agile beams locked onto their
respective signal paths.
FIG. 5 depicts a block diagram of the fixed beam search processor
according to an embodiment of the invention.
FIG. 6 depicts a block diagram of the agile beam search processor
according to an embodiment of the invention.
FIG. 7 depicts a block diagram of an agile beam element according
to an embodiment of the invention.
FIG. 8 illustrates the relation of a .SIGMA.-beam pattern with a
.DELTA.-beam pattern.
FIG. 9 illustrates the mono-pulse angle estimation curve according
to an embodiment of the invention.
DETAILED DESCRIPTION
Adaptive beam-forming can be implemented digitally in the
base-band. Therefore, some means of down-converting an RF signal to
base-band as well as a scheme for calibration is assumed. The
algorithm can be divided into two steps: 1) a low resolution search
performed by the fixed beam searcher, 2) high resolution tracking
which utilizes standard mono-pulse techniques. Alternatively, if
the approximate location of the mobile or fixed wireless user is
known or is determined by other means the low resolution track can
be circumvented with high resolution mono-pulse tracking only. The
simplicity of this direction-of-arrival estimation calculation will
enable real time tracking of a mobile station within the sector
area serviced by the antenna.
The low resolution search is performed by comparing the signals in
each of the beams that sub-divide the sector, (eg. 101-108). The
signals may be line of sight (LOS) or a time delayed non-LOS
multi-path. The base station equipment will decide which and how
many of the signals it will need, to meet the required
signal-plus-interference-to-noise ratio (SINR) for that service.
The base station will then cue multiple agile and autonomous high
resolution tracking beams to within +/-1/2 beam-width where each of
these signals reside. The agile beam is initially formed by phasing
the signals of each element in the array with a set of weighting
vectors corresponding to that fixed beam (101-108). With the agile
beam cued to the approximate location, the mono-pulse estimator
will determine the fine adjustment needed to lock on the signal.
The agile beam becomes autonomous and tracks the signal wherever it
is within the 120.degree. sector until it is handed off to another
sector or base station or the signal is lost. In a multi-path
environment with many obstructions, a signal from a mobile station
will likely become abruptly shadowed. Under such circumstances, the
tracking algorithm will lose the signal and its track. However, an
embodiment of this invention is the implementation of a low
resolution fixed beam search that can be running continuously in
the background, so that within the next timeslot or symbol another
signal will be acquired if one exists within the field-of-view of
the antenna.
Mono-pulse DOA is performed by measuring the relative signal levels
of two different beams formed with the same antenna. The summation
beam is formed by the summation of coherently phased signals from a
plurality of elements within an antenna array. The difference beam
is formed by symmetrically dividing the antenna and summing the
signals from one half of the elements, exactly out of phase with
the remaining half of the elements. The ratio of the difference
beam signal relative to the summation beam signal results in a
signal proportional to the magnitude of the angular offset and
whose polarity indicates to which side of the beam peak, the signal
is present. The tracking loop then tends to keep the maximum of the
antenna gain in the desired direction by keeping the signal
centered on the zero of the corresponding difference pattern.
For code-division multiple access (CDMA) systems it is assumed that
the RF signal has been matched filtered and de-modulated to a
digital base-band signal prior to any beam-forming. This assumption
includes the low resolution fixed beam search as well as for every
agile beam. Mono-pulse estimation therefore can be performed after
de-scrambling and de-spreading separates the desired signals from
other co-channel interference (CCI). This ensures that tracking
accuracy and precision is not degraded. For time-division multiple
access (TDMA) systems it is assumed that the radio-frequency (RF)
signal has also been matched filtered and delayed and that
synchronization is established. The beam-formed signal can then be
directly applied to the mono-pulse estimation circuitry.
Downlink beam-forming can be achieved by utilizing the DOA
parameters derived on the uplink. In FDD systems where the
separation of transmit and receive frequency bands are relatively
narrow it is known that DOA parameters are virtually invariant to
frequency. To reduce the probability that a single downlink signal
will undergo a deep fade, two or more signals can be transmitted
through multiple downlink beams derived from uplink DOA estimation.
The likelihood that the mobile will encounter a deep fade in two
different multi-paths is low. Therefore, it would be prudent to
transmit at least two signals in two downlink beams as disclosed in
U.S. patent application Ser. No. 09/987,722, filed Nov. 15, 2001,
and entitled "Method and Apparatus for Received Uplink-Signal Based
Adaptive Downlink Diversity Within a Communication System". The
uplink DOA information provided by this search and track method can
be directly applied to the downlink with the appropriate frequency
transformation of the array weighting vectors.
Angle estimation in one plane with an M-element linear array has
been considered above. This approach can be extended to 2-D DOA
estimation in both the azimuth and elevation planes utilizing a
2-dimensional (M.times.L)-element planar array symmetrically
divided into left-right and top-bottom halves. The signal from each
element is used to construct, not two beams as with the linear
array discussed above but a cluster of three beams that are slaved
to each other. The summation beam is formed by taking the sum of
all of the signals from each of the elements after they have been
phased properly to correspond with an appropriate steer angle which
lies within a specified maximum conical angle relative to the
normal of the plane of the array. The azimuth difference beam is
formed by summing the signals from every element after the same
phase gradient for summation beam has been applied to all of the
elements with an additional 180.degree. phase shift applied to the
elements comprising the left half of the array relative to the
elements comprising the right half of the array. The elevation
difference beam is formed by summing the signals from every element
after the same phase gradient for summation beam has been applied
to all of the elements with an additional 180.degree. phase shift
applied to the elements comprising the top half of the array
relative to the elements comprising the bottom half of the array.
This technique, when coupled with some means of time-of-arrival
estimation enables location based services.
The theoretical gains achieved by any smart antenna assume a
uniform distribution of users across a sector. If the density is
non-uniform, the higher density sub-sectors will suffer a
degradation in capacity improvement. However, multi-user detection
(MUD) schemes can also be integrated with the digital beamformer to
recover some of that loss. For example, if there are a cluster of
users within a small angular space it might be prudent to use one
or more of the fixed beams and some form of interference
cancellation normally associated with (MUD) to improve the
SINR.
It has been accepted that smart antennas offer an additional degree
of freedom for operators to improve the capacity and range of their
systems. The methods of this invention are much less complex, and
thus less costly than many of the other computationally intensive
eigen-decomposition techniques such as ESPRIT and MUSIC. It will
also perform far better than existing sectorization methods by
achieving a high gain antenna beam on every user while
simultaneously eliminating the need for frequent handovers.
Furthermore, the mono-pulse technique has been a proven method over
several decades in single beam Radar systems. Digital processing
technologies have now enabled such a technique to be used in the
multi-access, multi-user, fixed and mobile wireless communication
systems.
The adaptive beam-forming method comprises two functions: 1) low
resolution angle estimation performed by the fixed beam searcher
and 2) high resolution angle estimation performed by the agile beam
tracker.
FIG. 1 illustrates a multi-user multi-path path environment
illuminated by a set of fixed beams that sub-divide the sector.
Fixed beam searching is used for low resolution tracking and the
initial acquisition of a user. The fixed beam search (FBS)
processor segments the entire (120.degree.) sector 116 into M beams
(eg. 101-108), with cross-over points at a predetermined level X dB
118, usually 3 dB corresponding to the 3 dB beam-width of the
antenna.
Signals arriving at the base station antenna may arrive as
line-of-sight (LOS) or multi-path. Any one of the beams will
contain the signals from users within the main lobe of the beam as
well as attenuated versions of signals from users, outside of the
main beam. The attenuation levels are a function of the sidelobe
levels of the antenna pattern, 120,121,122,123. As shown in FIG. 1,
there are two users 113,114 that are within the field-of-view (FOV)
of the base station antenna 119. The signal 117 arriving at the
base station from 113 is a LOS signal incident from the direction
that beam 101 covers. The FBS will therefore assign that signal to
sector 101. The is no direct LOS path from 114 to the base station
antenna, however, two multi-path signals 109, 110 are present and
incident from directions covered by beams 108, 103, respectively.
The FBS will therefore assign the two signals to sectors 108 and
103.
Turning to FIG. 5, the M beams 101-108 are formed in the Fixed Beam
Searcher (FBS) 502 by performing a spatial transform of N complex
samples taken from each of the N elements at M (in this example
M=8) discrete points in angle space. Therefore, ##EQU1##
where, F(m) is complex and contains the total signal in beam m,
C(n) is the complex spatial sample from element n, W.sub.n,m is the
complex weight applied to element n for beam m, n is the number of
the element in the array from 1. . . N, m is the number of the beam
from 1. . . M.
The complex signal F(m) 505 from the m.sup.th beam is then
de-scrambled, de-spread or de-interleaved in the User Processing
Element (UPE) 506 for every user or multi-path signal k=1. . . K.
The UPE functions as the correlation receiver and separates users
from one another by de-spreading CDMA signals or de-interleaving
TDMA signals. The UPE does not necessarily have to decode the
polarity of the bit and therefore no channel phase estimation is
required. It is sufficient only to determine the magnitude of the
bit. A detector 507 for each user then determines the beam(s) with
the maximum signal level and integration may be performed to
increase the signal-to-noise prior to making the decision.
In the example, the FBS locates the signal from user 113 and
assigns an agile beam to it by passing the associated (N.times.1)
complex weight vector, m.sub.117.sup.fixed, 508 corresponding
directly with beam 101 to the agile beam search processor. Also in
the example, the FBS locates the signal direction of both
multi-path signals from user 114 and assigns two agile beams to
them by passing the associated set of compex weights,
m.sub.109.sup.fixed and m.sub.110.sup.fixed corresponding directly
with beams 108 and 103, respectively, to the agile beam search
processor.
The agile beam search processor contains at least K, agile beam
elements, (ABE), 602 (shown in FIG. 6). Initially, in each ABE, a
comparison is made between the current weights W.sup.i.sub.n,k and
the weights associated with the beam where the signal is located,
712. If the direction associated with the weight W.sup.i.sub.n,k is
within one beam-width of that found by the FBS than the agile beam
is tracking the signal and the beam becomes autonomous, otherwise
W.sup.i.sub.n,k is reset to m.sub.k.sup.fixed.
FIGS. 2-4 represent snapshots of beam positions as the agile beam
processor tracks and locks onto the several signals from the
several users. In the example, three agile beam elements, 602, are
assigned to two users. One user, 113 is assigned one ABE and the
other user 114 is assigned two ABE's. Prior to any information
about the DOA of each signal, the beam positions are arbitrary 201,
204, 207. Within the next symbol, slot, timeslot, or frame the beam
positions are cued to the established sectors 301, 304, 307
previously determined by the FBS. With the agile beams cued to the
correct sectors they become active and begin high resolution track
utilizing mono-pulse estimation.
FIG. 8 illustrates that a mono-pulse tracking system is implemented
by forming two beam patterns shown: the sum pattern 801 (shown
solid with a peak at 0.degree.) and the difference pattern 802
(shown dashed with a null at 0.degree.). Both of these patterns are
formed using the single antenna or aperture. The sum pattern is
used to provide absolute power information of the received signal
and is simply formed by summing the signals from several elements
of an array. The difference beam is formed by partitioning the
array into two halves phasing one half of the aperture to
180.degree. with respect to the other and then summing the two
signals from both halves. The ratio of the complex voltages
provided by the sum and difference beams provides the
discrimination information for DOA estimation and beam
tracking.
Referring to FIG. 9, it is seen that computing the real part of the
complex ratio of the difference pattern with respect to the sum
pattern over angle space, ##EQU2##
results in the mono-pulse angle estimation curve of 901. This
curve, representing the ratio of complex voltages at the sum and
difference beam signal ports, is proportional to the DOA of the
incident energy relative to the current beam position. For example
if the user is located at 0.degree., .DELTA.(.theta.)=0 and thus
.PSI.(.theta.)=0, depicting that tracking can be implemented with a
feedback loop that attempts to keep the beam centered at the zero
of the difference pattern.
FIG. 7 depicts a block diagram of the mono-pulse beam-former and
the tracking loop. Each ABE processes signals from a plurality of
antenna elements, N. The sum beam is formed in the .SIGMA.
beam-former 701 by performing the spatial transform as follows
##EQU3##
where, .SIGMA.(k) is complex and contains the total signal in beam
k, C(n) is the complex spatial sample from element n, W.sub.n,k is
the complex weight applied to element n for beam k, n is the number
of the element in the array from 1. . . N, k is the number of the
user or the user's signal.
The difference beam is formed in the A beam-former 702 by
performing the spatial transform as follows ##EQU4##
where, .DELTA.(k) is complex and contains the total signal in beam
k, C(n) is the complex spatial sample from element n, D.sub.n,k is
the complex weight applied to element n for beam k, n is the number
of the element in the array from 1. . . N, k is the number of the
user or the user's signal.
The .SIGMA. beam signal for a specified user, 703, is de-scrambled,
de-spread or de-interleaved in the .SIGMA. Channel Estimator 707.
Utilizing a pilot bit, the .SIGMA. channel estimator may also be
used to estimate signal phase, .phi. from the control channel for
W-CDMA systems. It can also deliver an integrated .SIGMA. signal
with higher SNR to the mono-pulse estimator. In W-CDMA systems the
handset modulator adjusts the gain of the control channel relative
to the traffic channel depending on the data rate. Higher data
rates require more signal energy to be stripped from the control
channel in order to maintain a specified bit-error-rate (BER). To
accurately estimate the channel phase, this integration may be
performed to raise the signal-to-noise ratio of the control
channel. The .DELTA.-demodulator, 706 performs the same
de-spreading, de-scrambling or de-interleaving and integration as
the .SIGMA.-channel estimator. However, if a phase estimate is
needed it will be provided by the .SIGMA.-channel estimator.
The angle estimation is performed by first taking the ratio of
signals from the sum and difference channels, as in Eq 1, 708a.
Then, by converting the dependent variable .psi. into an
independent variable and by converting the independent variable
.theta. into a dependent variable u=sin(.theta.), an expression can
be formed for the angle error relative to the difference beam
null,
The function .delta.u(.PSI.) is odd and can be approximated by a
Taylor polynomial,
where, .delta.u, is the angle prediction in sine space,
##EQU5##
Antenna patterns are a function of several variables including
aperture weighting, mutual coupling, array design, and beam
position. Thus, the mono-pulse estimation coefficients including,
a.sub.i will also be dependent on the same parameters. These
parameters are gathered by simulation and by measurement. Once
gathered they are stored in a look-up table 709 and retrieved. A
beam position correction can then be calculated, by using Eq 6, and
mapping it to a phase correction
for all of the elements, 710. This phase correction is then applied
to the current agile beam position, by recalculating the phase
weights
711, once it has been verified, 712, that the signal is still in
the same sector.
In the example the multi-path signal, 109 of FIG. 3 and FIG. 9 is
shown to be left of beam center. The ratio, .PSI. is then
calculated using Eq. 2, 708a the angle error is estimated using Eq.
6, 708b, which is mapped to a phase correction, 710 and then
applied to the current beam position using Eq. 8 and 9, 711. The
result is a new beam, 401 that is now centered on the signal.
Conclusion
A first method positions the highest antenna gain on the multiple
signal paths of a communication link between two or more
communication devices within a sector. The signal within the sector
is acquired and its location within the portion of the sector is
determined by utilizing a low resolution search comprising of a set
of fixed antenna beams that divide the associated sector into its
portions. By evaluating the optimal signal of those provided by
each fixed beam, the location is ascertained by mapping the beam
position to a portion of the sector. This mapping also produces a
set of antenna array weighting vectors for each and every
communication link within the sector of the communication system
associated with the antenna.
A secondary set of antenna beams is provided for each and every
desired signal path from the multiple users. The initial set of
antenna array weighting vectors are determined by the low
resolution search. This set of agile beams can be scanned
continuously within the sector while tracking the signals from the
users within the sector. Once initial acquisition is achieved agile
beam tracking is performed by mapping an angle offset determined
from the ratio of two signals. These signals are provided by two
slaved beams, the summation beam and the difference beam, that
comprise a single agile beam. The relationship of Equation 2
determines an amplitude that can be mapped as an angle offset
relative to the null of the difference beam using Equation 6. The
polarity of the result of Equation 2 determines which side of the
null the signal resides. Based on this offset a new antenna
weighting vector is calculated by equation 7, 8,and 9 and applied
to antenna elements. This continues at a rate consistent with a
timeslot (TDMA) or symbol (CDMA) until the user is handed of to
another cell or sector within the cell.
A second method positions the highest antenna gain on the signal
path of a communication link between two or more communication
devices within a sector. A primary set of antenna beams is provided
for each and every desired signal path from the multiple users. The
initial set of antenna array weighting vectors are predetermined
from a known user signal path location such as with fixed wireless
systems or from feedback information provided by a mobile user
about its approximate location within the sector or cell. A set of
agile beams can be allocated to each user and scanned continuously
within the sector while tracking the signals from the users within
the sector. Once initial acquisition is achieved agile beam
tracking is performed by mapping an angle offset determined from
the ratio of two signals. These signals are provided by two slaved
beams, the summation beam and the difference beam, that comprise a
single agile beam. The relationship of Equation 2 determines an
amplitude that can be mapped as an angle offset relative to the
null of the difference beam using Equation 6. The polarity of the
result of Equation 2 determines which side of the null the signal
resides. Based on this offset a new antenna weighting vector is
calculated by equation 7, 8,and 9 and applied to antenna elements.
This continues at a rate consistent with a timeslot (TDMA) or
symbol (CDMA) until the user is handed of to another cell or sector
within the cell.
* * * * *